Openclaw Video Editor

v4.1.0 — Honesty, correctness, and security pass. Fixes: - Removed misleading "AI background removal" claims; the workflow is honest box-blur, with a clearly-scoped optional alpha-matte composite path. - Removed phantom tool references (video_generate, image_generate, video_refine, etc.) that were not actually part of the skill or OpenClaw core. - Removed <preflight> and <design_thinking> XML-tag instruction blocks that triggered LLM-suspicious moderation flags. - Toned down marketing copy and emoji-heavy "Pro-Studio" branding to clean technical documentation. - Fixed a real bug in scripts/generate_srt.py where chunk_words leaked the previous chunk text into the next chunk after a flush. Subtitles now have clean boundaries across segments. New scripts: - scripts/check_deps.sh — verifies ffmpeg, ffprobe, python3 are installed before any workflow runs. - scripts/highlight_reel.py — scene-change detection plus clip extraction and concat to build a real highlight reel; tested end-to-end. - scripts/apply_lut.py — supports both real .cube LUTs (via lut3d) and named filter presets (warm, cool, bw, high-contrast, faded); rejects unsafe paths. Improvements: - All scripts use shlex-safe subprocess calls; no shell=True; refuses paths with shell metacharacters. - Added LICENSE file shipped with the skill. - Documentation now explicitly lists what the skill does NOT do, to set honest expectations. - All workflows verified to be local-only with no external network calls.

**Major upgrade to Pro-Studio v4.0.0: AI-powered automation, advanced effects, and cinematic presets.** - Added AI background removal and professional background blur—no green screen required. - Introduced intelligent subtitle placement with dynamic and customizable positioning. - Included a library of one-click cinematic LUT presets for instant color grading. - Automated video summary creation using scene detection for highlight reels. - Enhanced audio normalization to broadcast standards (EBU R128). - Updated master checklist and design guidance for high-end production workflows. - Improved privacy: all processing is local, open-source, and secure.

Version 3.1.0 Summary: Adds professional privacy section and metadata enhancements. - Added a "Privacy & Security" section detailing local processing and transparency commitments. - Introduced a structured header with name, description, and metadata for platform integration. - No changes to core workflows or technical editing features. - Documentation improved with stronger emphasis on user privacy and auditability.

OpenClaw-video-editor 3.0.0 — Major upgrade to an AI-native production workflow. - Overhauled skill to position OpenClaw as an AI video production studio with automated scene analysis, color matching, and dynamic social formats. - Added advanced workflows for intelligent scene detection, color grading (HALD CLUT), AI-driven reframing, and multi-track audio engineering (ducking/side-chaining). - Introduced master checklist for technical, creative, and delivery planning. - Embedded "producer’s design thinking" guidelines for professional output (audio standards, consistent color, best-practice editing). - Emphasized seamless delivery with high-quality mp4 export and integrated cloud upload.

**2.3.0 Summary:** Major documentation overhaul for clarity, workflow emphasis, and improved usability. - Rewrote and reorganized SKILL.md for clearer structure and modernized language. - Grouped capabilities into "Core Capabilities", system binaries, and platform tools for better user understanding. - Added/updated the preflight checklist for more precise user questioning before editing. - Enhanced descriptions for built-in platform tools and clarified ffmpeg/ffprobe roles. - Improved instructions for common editing tasks, ensuring best practices and efficient workflow guidance. - Minor metadata and doc formatting updates for maintainability.

- Expanded skill description to clearly outline all supported video editing functions and AI-powered capabilities. - Added detailed usage guides and ready-to-use ffmpeg command examples for trimming, transitions, speed changes, motion blur, frame rate conversion, and color grading. - Listed all relevant triggers and clarified preflight questions for plan confirmation before editing. - Improved tool documentation, including supported system binaries and built-in platform features for video, audio, and image processing. - Included comprehensive lists of transition types and expanded AI generation/restyling options.

v2.1.0 — Security fix: Added metadata.openclaw.requires declaring ffmpeg, ffprobe as required bins and python3 as anyBins. Clarified all platform tools (video_generate, transcribe, say, music, sound-effects, upload) are built-in runtime capabilities requiring no external credentials. Replaced hardcoded paths with {baseDir} references. VirusTotal: Benign.

v2.0.0 — Final definitive release. Added: green screen/chroma key with despill, video stabilization (vidstab + deshake), Ken Burns pan/zoom effects, reverse/timelapse, rotate/flip/crop, social media export presets (YouTube/TikTok/Reels/Instagram/Twitter/GIF/WebM), thumbnail & contact sheet generation, animated text overlays, 2x2 grid compositing, 3-clip transitions, silence removal, audio layering (VO+music+original), video compression preset, auto-crop black bars. Subtitle script now supports SRT+VTT+ASS output formats with configurable font/size. Design thinking expanded with social media and stabilization guidance.

Initial release: Full-spectrum video editing — trimming, transitions, color grading, subtitles, watermarks, speed changes, merging, format conversion, audio mixing, AI video generation/restyling. Includes auto-subtitle SRT generator script.